Solid state ultrasonic imaging

University of Warwick institutional repository:

http://go.warwick.ac.uk/wrap

A Thesis Submitted for the Degree of PhD at the University of Warwick

http://go.warwick.ac.uk/wrap/73983

This thesis is made available online and is protected by original copyright.

Please scroll down to view the document itself.

A thas i.s submitted by

S.O .. HARROLD, B.Sc.J~ng., A.C.G.I.

for the degr.ee of Doctor of l'hilosopby l.n

BEST COpy

AVAILABLE

TEXT IN ORIGINAL·

IS CLOSE TO THE

PAGE

. NUMBERS

CUTOFF

Abstract

The developm.e·ntand performance of a transistor scanned solid state ultra.sonic image conver-ter are described and the results of investigations into various applicable signal processing techniques t.o produce visible pictures are presented, The general requirements of an ultrasonic imaging system are considered and this is f'o11ml'ed by a critical survey of the more sensitive ultrasonic image conversion methods. The operation and basic performance of the transistor

scanned conver-ter are di scuased and its resolution and sensitivity Lnves ti.gat.ed , resulting in an apparent practi cal limit to the

reso-1

lution in. the image plane of 3'

mm,

and an estimated limiting sensi-tivity of approxima.tely 10-13 Wcm-2• Possible methods for the fabrication of a multielement converter are suggested and d.i.sous sed . Three sig!"ls.l process} ng techniques are investigated and compared; that using an acoustic lens in a conventional camer-a arrangement, an on-line computer which effectively performs electrically the f'unct.ion of El. lens, and a scanned array Doppler syst-em on which only a feas-ibility study was carried out. It is argued that if flexibility is important then the comput e r "lens" is the most fitting met.hcd , and

results for both near field focusing and far field beam deflection are presented) although at the moment.in I-dimension only , A relat-i vely s imp'Le extension to the pToe;ram will make 2-dimensional image reconst.ruc+i on possi.b.Le. The Doppler system descrd'bed pr'ovidea a simple eLectirica'l met.hod of determining the range and bearing of f'ar

Acknowledgement

-Theauthor is indebted to Dr. D.L Crecraft for his

understand-,ing and patient Gupervision, much of which has been carried out under

difficult circumstances, and also to Dr. H.V. Bhurmer- without whose invaluable support in the final stages this thesis would not have

been presented. Thanks are due to Dr. N.G. Meadowsand Mr. F. Dellow

for "their encouragement and assistance, and to Dr. D.S. Harris whose

helpful advice and interest in the last stages have provided the

necessary impetus to complete the work.

The author wishes to acknowledge the assistance of the Science

Research Council who financed the major part of the project, Lanchest.er

College of Technology (now Lanchester Polytechnic), and Portsmouth

Polytechnic, in particular the Department of Electrical and ElectronIc

Engineering, who supplied all the necessary facilities.

Special thanks are due to Mr. R.C. West whose except.Lonaf,

under--graduate project contributed greatly to the success of the work

des-cribed in Chapter

7.

Thanks are also due to Miss S.G. Westmore for

typing this thesis, and to the photographic section for reproducing

the Dlanyoscilloscope photographs, etc.

Above all the author wishes to thank his wife for her patience

Abstract.

Table of Contents

1

Acknowledgement. 2

6

List of figures, tables and programs.

Chapter 1. Chapter 2. 2.1 2.2 . 2.3

2.4

2.5 Chapter 3. 3.1 3.2 3.2.1 3.2.2 3.3 3.3.1 3.3.2

3.4

3.5 Chapter l~.

4.1

4.2

4.3

4.4

4.4.1

4.4.2

4.4.3

Introduction. 11

Ultrasonic Imaging Techniques. 15

Introductory Review. Electron Tube Converters.

Solid State Transistor Scanned Converters. Surface Aeons tic vlave Scanned Transducer. Discussion. 15 17 19 21 22

Signal Processing Methods ~lied to Ultrasonic Imaging.

24

Acoustic Lenses and Mirrors. Acoustic Holography.

Princip les.

Practical Considerations.

Image Reconstruction by Computer, General Principles.

Experimental Systems.

Scanned Array Doppler System.

Comparison of Processing Techniques.

24

25 25 28 29 29

30

32 33

The Solid State Converter. 36

Converter Development.

General Prindples of t.he FET Scanned Converter. Constructional Details.

The Transducer.

Choice of Transducer Material.

Source Resistance of Tuned Receiver Element. Acoustic Element to Element Cross-Coupling.

4.5 4.5.1 4.5.2 4.6 4.7 4.7.1 4.7.2 4.8

4.9

Chapter 5. 5.1 5.2 5.2.1 5.2.2 5.2.3 5.3 5.3.1 5.3.2 5.3.3 5.3.4 5. '. 5.5 5.6 Chapter 6. 6.1 6.2 6.2.1 6.2.2 6.2.3 6.3 6.3.1

The FET Sampling Gate. Static "on" Characteristics. Dynamic "off" Characteristics. Control Circuits.

Converter Performance. Sys tem Noise.

Sensitivity.

Future Development of the FET Scanned Solid State Converter.

Conclusions.

Ultrasonic Camera Using the FET Scanned Converter with an Acoustic Lens.

System Description. The Acoustic System. Transmitter Units. Acoustic Lens. Reflection Tests. Circuit Details. Timing Unit ,

Bandpass Amplifier.

Transmitter and Receiver Gates and Detector Unit. Staircase Generator.

Results. Discussion. Conclusions.

On-line Computer Lens.

Introduction.

Processing Principles. Focused Operation.

Response of a Focused Continuous Receiving Aperture to a Point SOUTce.

6.3.2 6.3.3

6.4

6.4.1

6.4.2 6.4.3 6.5 6.6

6.7

6.8 <-'hapter7.

7.1

7.2

7.3

7.3.1

7.3.2

7.3.3

7.4

1.5

References. Appendix

A.

Appendix B. Appendjx C. Appendix D. Appendix E. Appendix F. Programming Languages. Interfacing Principles. TI1e Experimental System. General Description. Phase Detector.

Analogue to Digital Converters. The Computer Program.

Results. Future Work. Conclusions.

Scanned Array Doppler System.

Introduction. System Principles. Experimental Work.

The Transducer Array aud Sampling Electronics. Measurement of the Characteristic Frequencies. Target Range and Bearing Tests.

Results.

Conclusions and Suggestions for Further Work.

General Conclusions.

Reprint of paper, "Solid state ultrasonic camera".

Ham.:.facturers' data.

Details of pulsed oscillator.

Reprint of letter, "Par-field sector scanning uS1ng a samp led multi-element ultrasonic rec.eLving array".

Reprint of paper to be published, "Control of

acoustic beam angle by harmonic analysis of sequential samples from a 1nulti·-ele.mentlinear ar ray'",

.Figure

1.1

3.2

4.1

4.2

4.3

4.4

4.5

4.6

4.7

4.8

4.9

List of Fig~res, Tables and Programs

A typical ultrasonic imaging system using an acoustic lens.

11

Formation of a hologram. 26

Image reconstruction by computer. 29

Pattern of electrodes on rear face of converter transducer.

37

Converter electrical system. 38

(a) Typical matrix output signal transmitter off. (b) Typical matrix output signal - transmitter on.

39

The FET scanning matrix before encapsulation. 41

Printed circuits for the F'ET scanning matrix. 42

The complete encapsulated solid state converter. 43

Transducer construction for 2nd stage of cross-coupling test.

48

FET scanning matrix circuit. 51

Typical low level drain characteristics for an n-channel junction FE'll.

5';

..

4.10 'l'estcircuit for voltage tr8l1sfer characteristic. 53

4.11 Voltage transfer characteristic for the gate circuit. 54

4.12 'rest circuit for gate breakthrough measurements. 55

4.13 Measurement of COFF• 57

4.14 Shift register control circuit. 58

4.15 Complete 100 element serial output of converter for 61

uniform "il1uminationfl•

l~.16 Equivalent circuit of transducer element, }i'ETand load. 62

5.1 Camera. system with acousti c lens. 68

5.3

5.4

5.5

5.6

5.7

Overall camera system block diagram. 70

Timing waveforms. 71

(a) Typical output of bandpass amplifier (pulsed mode).

(b) Output of detector for above input.

74

The ultrasonic transmitters. 76

(a) Reflections, with lens and test object positioned as for normal imaging. (b) Reflections, w:i.thlens and test object

removed.

80

5.8 Timing unit circuit diagram.

82

5.9 Bandpass amplifier circuit diagram. 84

5.10 Frequency response of bandpass amplifier. ₈₅

5.11

Transmitter gate circuit. ₈₅

5.12

Detector unit circuit diagram. ₈₈

5.13 Staircase generator circuit. 89

5.14

Staircase generator output waveform.

5.15

5.16

5.17

5.18

90

(a) Output signal for uniformly "illuminated" converter plate.

(b) Final pi.cture resulting from uniform "illumination".

92

(a) Converter serial output for half "shaded" plate.

(b) Final picture for half "shaded" .converter plate.

(c) Converter serial output for "unshaded" plate.

94

Image of 1·5 nm cross cut through 18 SWG aluminium sheet.

(a) Continuous wave transmission. (b) Pulsed transmission.

96

Image of 1'5 mm cross cut through 18 SWG aluminium sheet.

(a) Offset low.

(b) Offset high.

5.19

5.20

5.21

5.22

5.23

Image of 1'5 mm cross cut through 18 SWG aluminium sheet.

98

(a) Offset left. (b) Offset right.

Image of 1'5 mm cross cut through 18 SWG aluminium sheet, with arms at 450 to x-y converter matrix axes.

(a) Centrally placed.

(b) Offset so that one arm only imaged.

99

Image of 3 mm solid cross. (a) Magnification 1'9 times. (b) Magnification 1·4 times.

101

Image of 1'5 mm solid cross. (a) Using pulsed transmission. (b) Using pulsed reflection.

102

Image of two 2'S mm diameter boles through 16 SWG aluminium sheet. Spacing between adjacent edges

lot~

1·5 mm.

5.24 Image of two 3 mm electric spot welds. 104

5.25 Image of 10 mm diameter brass rod. 105

(a) . By pulsed transmission. (b) By pulsed reflection.

6.1 Path lengths between array and object plane. Ill.

6.2 Focused continuous aperture. 115

6.3 Input peripheral device selection and data transfer 122

logic.

6.4

6.5

6.6

6.7

Peripheral flag logic. 123

Outline of experimental system. 125

Phase detector additional logic. 128

(a) Reference and signal comparator amplifier circuit .:

129

6.8

6.9

ADC interconnections.

Simulated point source compared with real "point" source.

6.10

Computer reconstructed image of

10

mm transmitter.

6.11

Computer reconstructed image showing effect of maj or si de-lobe.

6.12

Focused near field images.

6.13

Focused near field lmages.

6.14

Images by reflection.

6.15

Perspex lens and computer "lens" comparison.

7.1

Basic principle of system.

7.2

Diagram showing how seq_uential sampling builds up the characteristic freq_uency.

7.7

7.8

7.9

7.10

7.11

7.12

Diagram used for derivation of theory.

Plot of f

Re

against target bea~ing.

Basic FET switch unit.

Experimental system for measurement of characteristic freq_uencies •

Experimental system for target range and bearing tests.

Positions of transmitters and the receiver array.

Output for transmitter No. 2 at 0°.

Output for transmitter No. 2 at 30•

Output for transmitter No. 2 at

7

0•

Output for transmitter No. 1at 0° and :No. 2 at

_7°.

7.13

Output using pulsed transmission. for transmitter No.1 at 00 end No.2 at

7°.

Results using pulsed transmission with transmitter No.1 at 0° and No.2 at 7°.

Upper trace - ,Output of r'esonarrt filter t.uned to fROo• Lower trace - Input to power amplifier.

7.15 As above but with resonant filter tuned to fR 7°.

7.16 Proposed system to provide a

PPI

display of target positions. Table 2.1 4.1 4.2 5·1 7.2 Program 6.1 6.2

6.3

6.4

List showing the more sensitive ultrasonic image converters.

Useful constants for the common transducer materials.

Cross-coupling test results.

Acoustic impedance and velocity for some common materials.

Experimental and predicted results for characteristic frequencies at s.sempiing frequency of 2 '26 MHz.

Experimental and predicted resUlts fer characteristic frequencies at a sampling frequency of 4'52 MHz.

Harmonic analysis of output signal.

Characteristic frequencies for ultrasonic frequencies of 547 kHz and 528 kHz.

Results for target 1 at 7°, target 2.at 10°.

Results for target 1at 0°

,

target 2 at 10°.

Computation for path length phase increments.

Image reconstruction routine.

Phase and amplitude data acquisition.

FOCAL image ;reconstruction for simulated point. source.

1.. Introduction

Imaging, as distinct from other fOrm£ of information display, may be defined as being the formation of a visible image of an object plane or space, the latter lying usually norma1'to the axis of wave propagation and parallel to the converter image plate. The image has height and breadth only, there being no time related dimension except in the special

case of 3 dimensional imaging. Thus the photographic camera, television, radiography and the eye are all imaging systems • A basic ul trasonic imaging system is shown in figure 1.1 in which it is assumed, as is nor-raally the case, that conversion to an electrical signal is performed as an intermediate step between the ultrasonic and optical images. Parts T,

o

L

~==B~;:---t--l+'i~

L ~--~

o

A typical ultrasonic imaging system using an acoustic lens.

O; Land C are all innnersed in a liquid (usually ware r) so that they are acoustically coupled. A generator G produces an electrical sinusoi.dal drive signal at the appropriate ultrasonic frequency to the

electro-ac.QUS tic transmitting tr ansduce r T which radiates towards the test object O_• ~$'~J~te. "'~1\'3_{u~££raet~9n} d . ... d . 1" • h

object and the acoustic lens L focuses the diffracted rays to form a

diffraction image on the image plate of the acousto~electric converter

e. The output of C is normally in the form of a serial line-by-line

scan of the image plate, being similar to a television video signal.

After amplification by A and, if necessary, detecti.on by D it is

pre-sented to a eRO as an intensity modulating signal. Scanning of both

converter plate and eRO is controlled by unit S. This system uses

through transmission, the transmitter being sited on the opposite side

of the object to the receiving converter. Sometimes it is advantageous

to have the transmitter placed on the same side as the converter so that

the latter collects reflected signals from the test object.

The possibilities of ultrasonic imaging were first pl't1ctically

1

realised by Sokolov in 1937. He successfully produced an €.lectron

scanned piezoelectric plate "camera" system which effectively became the

• 2, 3 I.

prototyp? for development by other workers in RUSSla , Germany s

5,6 7

Britain and the

U.S.A.

over many years. Although improvements arE'

still being made, these converters possess some serious inherent

short-comings, as is evi dent when it is realised that they are still ill the

main used only.as laboratory research instruments.

The requirements of a flexible ultrasonic image conversion system

may be listed as follows:

(i) High sensitivity.

(ii) Image plate dimensions equivalent to at least

100 x 100 vave Iengths in the transmission medium.

(iii) Linear resolution equal to or better than 1

wave-length in thetransmi.ssion medium.

(iv) Retention in the output electrical signal of the

(v) Pulsed operation possible.

(vi) High reliability and minimal setting up

procedure.

The objectives of this research progrannne may be-divided into two

clear sections:

(a) The production of an ultrasonic image converter to satisfy

as many of the above requirements as possible, and,

(b) To investigate and compare various ways of processing the

signals to produce a visible image using the converter

developed in (a).

Chapter 4 deals in detail with (a) and chapters 5, 6 and 7 with (b). The

objective of section (a) h?s been achieved by effectively fonning the

converter image plate into a matrix of piezoelectric elements which are

then sequentially sampled by arrays of field effect transistors. The

result is a highly sensitive and reliable system with potentially none of

the inherent limitations of the Sokolov type tube converters. Three

forms of signal processing have been considered for section (b), an

acou-stic lens, a form of holographic reconstruction using an on-line computer

and a scanned ar ray technique based on the change in Doppler frequency

shift associated with target bearing. The techniques are compared and

their limitations and those of the solid state image converter discussed,

together with suggestions for further work, in chapter 8.

Ultrasonic imaging has important applications in two widely difierent

_areas, underwat er visualisation for ranges of up to 10 m or so, and

medi-cal diagnosis and associated work. Television is preferable for

under-water viewing in clear water, but most requi.rementis are for coastal or

conditions ultrasonic wave propagation, although suffering increased

30

attenuation, is still possible Medical uses would include amongst many

others, observation of the foetus, where X-rays are damaging, and heart

surgery where continuous visualisation is .an advantage. Short pulsed

31

operation is essential ,however, if the image converter Ls ever to

be-come a useful medical tool. Non-destructive flaw detection of solid

mate-rials, although a very desirable and much sort after application, is

inhe-rently a more difficult problem. The ima.ge converter requires virtually

perfect homogeneous coupli.ng between its whole frontal surface and the

object to be viewed. In the unlikely situation where this has an

appro-priate and large enough flat surface, then coupling may be achieved

direc-tly using a thick oil or grease, but for most applications immersion in a

coupling fluid (water or oil) is necessary. This introduces a reduction

in the system sensitivity due to impedance mismatch, and loss of defillition

due to ref raction caused by the irregular or non-coplanar face of the

ob-ject through which the ultrasound passes.

It is evident that there is an as yet unsatisfied need, over a very

, wide field, for a suitable ultrasonic image converter. It is hoped that

the FET scanned solid state converter may, after further development, prove

2. Ultrasonic Imaging Techniques

2.1 Introductory Review

Although many forms of ultrasonic image converters have been

devised, all but the electron tube and solid state types would appear

to be either too slow or to possess insufficient sensitivity. Table 2.1

lists for comparison some of the more sensitive forms of converter that

have been developed, giving their measured sensitivities, resolutions

and image formation times. The list has been restricted to those wi.th

-3 2

sensitivities greater than 10 W/cm. The first two provide direct

conversion to a visual image whereas the rest are all acousto-electric

and therefore require some form of electro-optical converter in order

to complete the imaging process. There are other solid state conve rte rs

whose send tivi ties have not as yet been pub lished and are t'he'ref or'e not

included in table 2.1.. Details of these are given in sections 2.3 and

2.4.

There have been several comprehensive reviews of ultrasonic imaging 12, 13, 1~

methods This review will be limited to electron tube and solid

state converters since the former is the only type with a comparable

Technique Limiting sensitivity W/cm Resolution Image formation Time. Seconds.

Optical detection of liquid surface

deformations.

Mechanical alignment of flakes in liquid -Pohlman CeU9•

Soko1ov tube:-Quartz face p1ate6• Ba Ti 03 face p1ate3

Cathode stabilised

tube:-Quartz face plateS.

Mechanically scanned quartz plate10•

Solid

state:-Bipolar transistor scanned PZT p1ate11

(20S MHz)

FET scanned PZT plate (the product of this

research programme)

-3

10Sx10 A few mm < 1

2°Sx10-3 several wavelengths < 1

Depends upon intensity

-7

1\

7xlO -S

I'

103xlO

-7

104x10

3x10-12 )

1 + 2 times thickness of plate.

,1\

loS wavelengths j < 002

)

< ·02

)

I

3 + 120

I

I)

2.2 Electron Tube Converters

Sokolov in 1937 pioneered the ultrasound camera with his ultrasonic 1,15

microscope ,although he gave very few details at the time. He was

the first to realise that a liquid damped. piezoelectric plate, rather

than vibrate as a whole, responds point by point to the incident local

sound intensity. This fact was vital to the satisfactory operation of

the image converter, which used a single homogeneous piezoelectric plate

as the first stage in the co.nversio.nprocess. Cameras later develo.ped

2, 3 It 6 7

in Russia , Germany ,Britain and America are all based on the same

principle of operation. A piezoelectric plate forms the end wall of a

cathode ray electron tube· so that its back face may be scanned by an

electron beam, whilst its front face is subjected to the ultrasound

pressure image. At equilibrium, the piezoelectric plate being a geod

insulator, the primary e1ectro.n beam current will equal the secondar.y

electron current passing back to what is termed the collector. However,

the ultrasonic pressure variations on the front of the plate change the

voltage on the back face, and the equilibrium voltage therefore requires

reestablishing~ This means that the secondary electron current will

reduce as the piezo voltage tends to rise and increase as it tends to

fal1. The output from the collector, consequently, takes the form of

an amplitude modulated wave at the ultrasonic frequency, the mo.dulation

representing the line-by-line spatial distribution of ultrasound Lnt

err-si.ty across the piezo plate. In order to produce a visible image this

electrical signal may be fed to the grid of a conventional display

cathode ray tube whose scanning is synchronised to. that of the converter.

This form of e1ec.tron scanned tube is termed collector stabilised.

5,16 17,18

Smyth and Sa.yers between 1954 and the early 1960's

which Sokolov had published on his tube. Its operation differs in one

major respect; the output is not coupled to the piezoelectric plate

voltage via a secondary electron current. Instead a signal plate is

included on the outer surface of the transducer and there is no

collec-tor ring within the tube. Acoustic to electric conversion is

effecti-vely completed in two separate electron scans. First, with the

ultra-sound off, a low velocity primary electron beam scans the target so that

no secondary emission takes place. Conse.quently, in order to satisfy

the equilibrium conditions, the target surface potential will be reduced

to that of the cathode at which stage no further electrons will reach

the target. The ultrasound is then switched on, a pressure image focused

onto the front face of the transducer, and the target once more scanned.

More electrons may now arrive at every point on the target surface where

the ultrasound impinges, but only during the positive half cycles of the

piezoelectric potential excursions. A charge distdbution is thus built.

up over the surface of the target which will be directly proportional to

the ultrasound intensity at any point. The build up of this charge ld11

induce corresponding currents in the signal plate, which is earthed to

the ultrasonic frequency, and these are used to provide an electrical

output signal. With the ultrasound off the first electron scan is then

repeated and now performs a discharging process in order to remove the

negative charge distributed over the target face.· The low velocity

electron beam contains a sufficient number of positive ions to make this

possible, and consequently the target potential is raised to that of the

cathode in the time available. The target is then ready for the next

charging or imaging scan. The output of this converter is a serial video

signal corresponding to the line-by-line distribution of sound intensity

the Soko10v tube. Because the target potential is maintained-at that of

the cathode, this type of tube converter is known as cathode stabilised.

2.3 Solid State Transistor Scanned Converters

Two bipolar transistor scanned ultrasonic image converters were

developed quite independently between 1966

11

under the direction of Knollman ,and the

and 1968, one in America

19

other in Britain, forming

the early pa-rt of this research programme (see Appendix A).

The former consisted of a 2·5 MHz half-wave thickness resonant

piezoelectric plate having on its uninsonHied face a 10 x 10 matrix of

square silver electrodes, and with overall dimensions of approximately

0·7 x 0·7 inches.

A

10 x 10 matrix of transducer gates controlled in

rows and columns effectively scans the transducer element s and connect them sequentially to a common load. In order to overcome the capacitive

signa.1 feedthrough in the "off" gates at 2'5 MHz, frequency conversion

is carried out within the "on" gates creating a relatively low (500 ~Hz)

intermediate frequency. By selectively amplifying only this Lntermediat.e frequency, the feedthrough signals which will be at 2·5 HHz are rejected.

Mixing is performed by interrupting the gating signa.ls at the local

osci-Llat.or frequency so that a time varying switell conductance results. The sensitivity of thds vconverte'rbased on initial tests was approximately

-12 20,21

5 x 10 W/cm2, but later publications claimed figures of around

10-11 and better than 10-10 W/cm2•

The transistor scanned converter developed here at the start of

this

project, although basically ve.ry similar to the couver t.er described above ,

did not employ mixing to reduce the cross+coup ling problems. For this

reason it was limited to operating frequencies of 1 HHz or less. A

compo-site clamp-series switch a.:rr~~llgenent was investigated in an attempt to

10 mV or so. Since the clamp transistor had an inherent offset voltage,

signals of lesser amplitude than this were not affected and the circuit

therefore offered little advantage over that using a single series

transistor. Matrix drive signals were derived from two shift registers

rather than decade counters and decoders. Results were marred by

unreliable connections to the transducer elements and by the high level

of cross-talk or breakthrough signal (typically 50% of the total output

sign.al).

No

meaningful sensitivity tests were therefore made, and

further work on this converter was quickly abandoned in favour of one

using FET gates. These gates showed great promise, with an "on" to "off"

signal output ratio of greater than 50 dB's at 6 MHz compared with less

than 30 dB's at 1 MHz using the bipolar transistor gate. This lead to.

the development of the FET scanned piezoelectric plate image converter

which has formed the basis of most of the experimental work carried out

under this research programme.

A detailed description of the FET scanned converter is given in

chapter 4. Operating at an ultrasonic frequency of 1 MHz it uses a

PZT.5A ceramic disc transducer, the rear face of which is formed into a

10 x 10 matrix of separate square silvered areas. Each of these areas

is connected to a common output line via two FET linear gates, one t.o

locate the row and the other the column. Two shift registers control

the gates so that the transducer elements are scanned sequentially line

by line. The output thus takes the form of an amplitude modulated wave

at the ultrasonic frequ.ency, the modulation representing the ultrasound

intensity distribution across the front face of the trrulsduceT. For

unity signal to noise ratio at the electrical output, an ultrasound -13

22

The principle has been further developed by Maginness at the

University of Stanford in 1973. A transmit-receive mode of operation

at an ultrasonic frequency of 3-5 MHz is used in conjunction with an

acoustic lens. The scanning both for transmitting an? receiving employs

special FET integrated MOS arrays. These are bonded directly to a thin

interconnecting layer which carries the necessary address and signal

conductors and wh:i.chis buHt up on the back face of the transducer

mat-rix. The result is an extremely compact and reliable arrangement, and

although at the time only a 10 x 10matrix had been produced, extension

to a 32 x 32 was considered quite feasible. The production of such a

device using modern solid state technology has been the long term aim

of this project (see Appendix A). Realisation of this objective has

been prevented by the problem associated with reliably bon-l i.ng the many required electrical connections on the underside of a silicon chip to

the appropriate signal and addressing conductors on the back face of the

transducer. With this overcome the solid state ultrasonic image

conver-ter should at last become the viable tool which to date no other

ultra-sonic image converter may claim to be.

2.4 Surface Acoustic Wave Scanned Transducer

23

A new type of ultrasonic imaging device which uses a-coustic

sur-face wave delay lines in.conjunction with FET mix~rs to scan a trans-23

ducer matrix in two dimensions was proposed by Havlice et a~. at the

University of Stanford in 1973. A one dimensional version had at that

stage been constructed, and, in conjunction with mechanical scanning in

the other dimension, produced. acceptable visual images on a eRO. The

single delay line has taps evenly distributed along its length, each tap

signal, at the ultrasonic frequency 00 , from the appropriate transducer

s

element. The mixer outputs are commoned so that as a short pulse of

frequency 001· is sent up the delay line and successively passes the taps,

transducer elements are effectively scanned to produce a serial output

at an intermediate frequency of 00 + 001' In addition to linear scanning

s

it is relatively easy by using a suitable frequency modulated linear

chirp as the input to the delay line to create a focusing action and

hence remove the necessity for an acoustic lens. '£Wo dimensional

scan-ning was to be achieved by using two acoustic delay lines, one to scan

the x direction and the other the y, forming a rastor scan of the

trans-ducer elements which lies at 450 to the x and y axies.

2.5 Discussion

The primary requirement of the ultrasonic image converter, in order

that the signal processing detailed in chapter 6 might be carried out,

was that it should retain in its electrical output the phase of the ultra-·

sonic signals. The Smyth cathode stabilised tube did not fulfil thi,s ,

and at the beginning of the research programme the Sokolov collector

7 24

stabilised tubes improved by Jacobs and Kennedy were not yet

commer-cially available. In any case these tubes which included an electron

multiplier, although more reli.able, did not have a significantly improved

general perform.ance. The inherent limitations which are brought about by

the necessity for a vacuum and the fact that the signal path includes a

low conductivity secondary electron current section were still present.

The former limits the maximum safe image plate diameter, although there

24-29

have been numerous attempts to overcome this , and the latter limits

the sensitivity due to both its noise content and the attenuation it

A new form of image converter, therefore, was required and with the concept of the solid state transistor scanned transducer it became a

worthwhile aim in itself. This converter, later developed to an FET

scanned version (see chapter 4), does not·have an inherent limitation on

image plate diameter, although in a non-integrated form the cost becomes

a prohibitive factor. Sensitivity measurements indicate this to be

possibly 5 orders higher than that claimed experimentally by Sokolov.

This comparison, however, takes no account of the "noisell introduced by

the transducer element to element output variations, which in the present

converter for even image plate insonification may be as high as 20%.

Image resolution across the piezoelectric image plate could be somewhat

better, since the transducer's rear face may be acoustically damped.t.o

further reduce the mechanical

Q.

The solid state converter developed by

22

Maginness has greatly improved the prospects of this form of convert~r,

since its unique transmit-receive capability may be used with both lenses

and electrical signa.l processing techniques. The surface acoustic wave

delay line scanned converter by Havlice, a completely new method, although

23

up to the present only briefly described ,could prove to be an extremely

flexible and efficient arrangement, providing the proposed 2 di.mensional

scanning is practically feasible. Capacitive cross-coupling between the

effective x and y strip electrodes placed on each side of the transducer

3. Signal Processing Methods Applied to Ultrasonic Imaging

3.1 Acoustic Lenses and Mirrors

Acoustic waves in almost all respects behave in the same manner as

light waves; they are refracted, diffracted, reflected and absorbed, but

show little dispersion. There is an important ~ifference however;

acous-tic waves are subject to mode changes, particularly at liquid solid

inter-faces, where a proportion of the longitudinal wave energy (if that is the

original mode of transmission) is converted into shear waves (there is

also some conversion to surface waves). These increase as the angle of

incidence of the longitudinal waves to the interface increases.

On the basis of refraction and reflection it is clearly possible to

focus acoustic waves using both acoustic lenses or mirrors, and to thereby

form acoustic images. The resolvable detail of these will, in the main,

be limited by the same criteria as those applicable to optical Lmages .

Irr~ge aberrations will occur, particularly if a simple spherical design

is used, and the best achievable resolution in the image plane will be the

Airy disc of diameter 1·12

A

!,

where

A

is the wavelength in water (the

transmission medium), f is the focal length and d the diameter of the lens

or mirror.

The material of an acoustic lens should have an acoustic i.mpedance as

near as possible to that of the surrounding medIum in order to minimise

reflections) and should have a significantly different acoustic

longitu-dinal wave velocity. The r'eEl.ected energy increases with the angle of incidence, and therefore it is advantageous for the lens material to have

a. high refractive Lndex, since this results in a larger radius of curvature for a given focal length. The processing action performed by a convergent

lens may be considered as being a spadal spherical phase advance function

greater phase advance than those on the axis. This is achieved by formlng

the lens into a spherical shape such that, in the case of a lens material

where the acoustic velocity is higher than the surrounding medium, it is

concave, and for the reverse situation it is convex.

Acoustic mirrors have the advantage that complete reflection is ideal

and in practice is more easily achieved than the perfect impedance match

required by a lens. Where very large apertures are acceptable, for examr 32

pIe underwater imaging ,mirrors are very successful, but for the

visual-isation of small objects at short range a short focal length is needed,

resulting in an aperture limitat.ion. In the latter case, the converter

may seriously obscure the mirror, since the mirror is necessarily on the

same side as the incident waves to the converter. The problem becomes

clearly impossible if unity magnificatio~ is r.equired; the object and image

planes are physically in the same place. The signal processing action of

the spherical mirror is almost identical to that of the spherical lens;

non-paraxial rays are reflected before paraxial and a spatial sphe rical

phase advance function thus operates on the reflected rays so that they

are brought to a focus in the image plane.

3.2 Acoustic Holography 33,35 Principles 3.2.1

In order to establish a simple physical pict ure of the principles

involved in the formation of acoustic holograms and their image reccns

truc-tion, a 2-dimensional system will be considered using geometric ray theory.

No loss of generality results from the consideration of only t\V'0 dimensions.

Consider a ray, defined as a narrow beam which nevertheless is a

at z

=

z and meeting the Z-axis at z , as shown in figure 3.1.

o

PLANE.

Re.FE.RENCE. SE-AM

z

Acovs rtc RAYS

'- HOL.OGRAPH IC. PLAKe:..

Figure 3.1 Formation of a hologram.

In the region of zl this wave may be represented by the expression:

A(zo) cos

[wt

+ Hzo) +

2:

sin • zJ ...

3.1

where

w

and A are the acoustic angular frequency and wavelength r

espec-tively, t the time, ex: _{the angle the r ay makes with the X-'axis, A(z ) the}

0

amplitude of the ray, z a point on the Z-axis near zb and

q,

(zo) t.he

phase of the ray. Similarly the reference beam may be represented at the

Z-axis, assuming unity amplitude, by the expression:

cos

[wt

- T

2w s m• 3.2

where 0 is the angle of the reference beam to the x·-axis. Now consider

S\l'W\'MQd.J

• •• • o.'II\Q

the effect of tbese two waves be1ng malt1~11eft together on the Z-aX1S ~ $~u.fY\I::I!t ~u. ..TG.d.

9R t1.::te~=plaQ9. The relevant term of the result would be:

Note that this expression is independent of time and may thus be recorded

for various values of z to form an intensity pattern or hologram.

Assuming that the hologram has been recorded· on an optical

trans-parency, image recons t.ruct i.on is carried out by means .of a beam of coherent light, and if this is considered to be plane, then, in the plane of the

If

hologram may be represented by the expression:

cos

[WI

t - (~:)(~) sin 0.z ]

3.4

where

wI

and

Al

are respectively the angular frequency and wavelength of

the Li.ght , The inclination of the light beam to the X-axis is a function

of the ratio of the optical to acoustica'lwavelengths, and is given by:

sin-I

[(¥)

sin

e ]

...

3.5

It is therefore approximately the angle

e

reduced by the ratio of the wavelengths.

The rays passing through the hologram will have their amplitudes

multiplied by the intensity patterns recorded by the origina.l object rays.

For the light rays passing t.hrough the region around Zl of the hologram the wavefront emerging is obtained by multiplying together expressions 3.3

and

3.4,

resulting in:

1

A(z ) cos

o [ Wj t + ~(Z ₀) + (27T)

_Al

(4t)

sin u:.z

]

[ wI t

- Hz )

(27T) (b.) (2 sin

e

+ sin

")J...

3 '.0

0

Al

A

+

1

A(z ) cos

o

Comparing the first term of expression 3.6 with expression 3.1 it will be

seen that they are identical except that the former is inclined to the

X-axis, or hologram norma}, at an angle of:

3.7

A

at an approximate distance of (-)x from the hologram, where x , as 'vill

Al 0 0

be seen in figure 3.1, is the original object Z-axis distance. Similarly

\

the second term of expression 3.6 represents a real image of the object

at the same distance from the hologram as, the virtual image, but angularly

separated.

3.2.2 Practical Considerations

33, 34, 36

Most successful experimental systems use an electrical square

law detector or multiplier to perform the essential reference signal

multiplication. If a multiplier is used then it is possible to introduce

the reference as an electrical signal whose phase is automatically

control-led by the mechanical scanning of the Z-plane by the a.coustic probe. At

present almost all Z-plane sensing is carried out by using a mechanically

controlled probe which scans the required'area at a suitable pitch. The

application of an electrical/electronic acousto-electric image converter

to this problem would offer a tremendous advantage, but the resolution

required, and in particular the aperture, is too great for any device

available at the moment. In order to correct for the vast difference in

the wavelengths of the acoustical and optical waves the hologram i.s

reduced by a. certain factor (around 10), and whilst this reduces thein~ge size by the same factor, the distance that the image lies from the

ho10-gram is reduced by the square of this factor. rrhi.~effect may be corrected

in practice by optically magnifying the image before viewing. 33

Aldridge used in his system 1 llS pulses and an ultrasonic frequency

1

of 10MHz, and this gave him a limiting resolution of about

6

mm in water and, with range gating, provided a resolution in depth of around 1·5 mm,

A facsimile recorder synchronised "to the probe scanner produced the

ori-ginal i.ntensity plot, and this was then photographed and reduced in size 37

used in conjunction with a computer to perform simulated holographic

functions and it is claimed that the process is superior to the normal 38

area holograms. The application of a scanned linear array has been

considered, where a transistor scanned linear array of transducers is

moved mechanically to create the second dimension.

The striking 3-dimensiona1 visuaHsation afforded with optical

holo-grams is not in practice possible with acoustic holograms. Because of

the effectively very large apertures readily attainable with light, the

viewer may look at the image through only a small part of the aperture

and in so doing may view the image from different directions. In

prin-ci.pLe the same is possible in the acoustic case, bu.t in practice the

apertures are sma11 (a relatively small number of wavelengths across) and

the loss of resolution resulting from using only a small portion of the

aperture is usually unacceptable.

3.3 Image Recons.truction ~ Computer

3.3.1 General Principles

Figure 3.2 shows the basis of an imaging system comprised of an

electro-acoustic image converter and an insonified object situated in the

object plane. The 8.rrangement is 1ensless, the computer providing the

necessary signal processing to image the object.

£.L.e.CTRO-AeOIJSTIc:. COM PU'T ER .--\

r

C.ONVG.RTE.R •

r---...;:..-- ...~

oSJe:.c,:r

r

PL.f,NE..

p

IMAG-E.

AMPL.fTI.IOE DELAYS OF POINT ooC1l-¥---- ,\1'10

P, SUMIV.A'rlol'll.

The action of a lens, as described previously, would be to focus

the divergent IIrayslleminating from the object so that only at one point

in the image plane was the energy from this source concentrated. It may

be considered to do this by imposing a spherical phas~or time delay

function on the incident "raysll so that, as they leave, their phases are

such that the resulting signals are only in phase at one point, the focus.

There are thus three regions where spatial time variations occur; between

object and lens, within the lens itself, and between lens and image.

The computer must provide the last two in order to perform the analogous

function. It must arrange that the phases of all the received signals

radiating from the object point are equalised before a summation is made

to produce the image magnitude of the point. The computer delay function

for a given situation, although apprQltimately spherical, is therefore not identical to that of the lens.

3.3.2 Experimental Systems

A system using very short pulses and in consequence possessing a·wide 39,41,42

bandwidth was developed by Maginness , and this resulted in

provi-ding the additional facility of being able to Image over a range of harmond>

cally related frequencies. Transmitted signals consisted of pulses of

approximately 2 cycles duration of a 5 HHz wave, and recordings were made

of the signals reaching the receiving aperture by means of a mechanically

scanned transducer matrix. The signals were sampled at an effective rate

of 24 MHz before quantisation and conversion into a paper tape format.

After completing the recording of all 144 matrix element signals, the

tape:; were fed into a large computer, programmed to perform the necessary

processing. The latter took the form of a fast. Fourier transform. The

output, in the form of a 2-dimensional graded intensity print out,

slices through the material being viewed (an aluminium block) by calling

up the various object planes and frequencies as desired. The result was

a very satisfactory and flexible processing system, the l~eak link being

the long time involved in both the mecharri.cal scanning of the transducer

matrix and the production of and inputting to the computer of the paper

tape records. 40

Shondhi developed an experimental system in air at an audio

fre-quency of 6·7 kHz, again using mechanical scanning and off-line computer.

Results, showing the images of simple objects, are impressive.

Both the systems described suffer from the one major disadvantage;

the long delay between the acoustic image formation and the production of

the final visual image. The wide-band system, if it was to use its

wide-band properties, had a further problem; it requir.ed very large immediate

access computer word storage. These two factors encouraged the

investi-gation into an on-line narrow band single frequency system, described in

detail in chapter 6, and which forms one of the main applications of the

solid state ultrasonic image converter. For most anticipated practical

applications of any imaging system, real time working is almost essential,

a few seconds being at the most acceptable. Thus an essentially simple

approach was considered in order to minimise computer time and storage

requirements. Hence the processing performed by the computer is basically

as outlined in section 3.3.1 and consists of the addition of appropriate

phase increments to the digitised transducer element signals so that on

summation they appear to eminate only from the point in the object plane

being imaged. The elements of a 10 x 10 Uk~trix are scanned electronically

(see chapter 4) and the phase and amplitude of each measured, converted

into a digital form, and transferred to the computer store within the.

the computer modifies the phases of the element signals to "focus" a

particular point in the object plane. The modified signals are summed

to complete the "focusing" action for the point, and a repetition of this

process for all the other points in the oQject plane produces, after

suitable digital to analogue conversion, a visual picture.

3.4 Scanned Array Doppler System

An echo. returning from an insonified target situated in the far fie.ld

of an array of ultrasonic transducers will strike this array at an angle

dependent on the target bearing. If the array is linear and consists of

equally spaced elements which are sampled sequentially at a known and fast

enough rate, then the resultant electrical samples after demodulation will

produce a sinusoidal signal whose frequency is a function of the angular

bearing,

e,

of the target. This resultant frequency is given by the

Doppler expression:

[

C

]

fR fT 1

s

sin

e

:.:

Cl

where C

=

f d is the sampling velocity across the array,

s s

3.8

fT is the frequency of the incident wave,

d the distance between adjacent element centres,

Cl the velocity of sound in the transmission medium,

and f the sampling frequency. s

In order to satisfy the sampling theorem f must clearly always be

s

greater than 2fR thus giving a boundary to the range of

e

for a given is'

In practice with a finite array, and where sampling is continuous,

only certain discrete characteristic frequencies corresponding to discrete

For a continuous resultant sine wave to be produced the phase of the

signal from the last e.1ement should be l1e1>behind the phase of the first,

where 11~ is the phase difference between adjacent elements. The discrete

bearings caused by this limitation are given by:

e

=

S1n• -1

3.9

~lere q is an integer,

and n the total number of elements.

The Hmi tat ion expressed by this equation is brought about only by

the manner in which information is extracted from the resulting signal.

A s~ngle set of n samples across the array contains all the required

bearing information.

43, 44

A practical system was developed to establish the validity of

the theory and details of this will be found in chapter 7. It proved to

be a successful preliminary investigation and showed that further more.

detailed work would be useful. Although not: referred to at the time a

48

similar scheme had been attempted in 1962 using a scanned line hydrophone

The electronics used was not at the time good enough, it would seem, to

encourage further work, and although some useful results were obtained

particularly regarding noise performance, no publications since that date

have, as far as is known, been produced.

3.5 Comparison of Proc~ssing Techniques

The use of the signal phase infonnation forms a common thread which

runs through all the t't:chniquesdiscussed above. In the case of lenses

beam, and for the scanned array Doppler system the instantaneous amplitude

of the incident wave is sampled and the samples analysed for a phase

coherence at certain frequencies.

By far the simplest processing method is that using lenses or mirrors.

If carefully designed these can produce results which are difficult to

improve upon by other processing techniques using comparable aperture

dimensions. They are, however, somewhat inflexible in that the field of

view and magnification may not be rapidly changed when desired. They

require an intermediate coupling medium between the converter and the

object to be viewed, and whereas in an immersion system this presents no

problems, where imaging within the human body or within relatively large

solid objects is required, direct coupling of the converter to the body

is advantageous. With the other three processing methods discussed this

is possible, although it is not suggested that this could be usefully

applied to the scanned array Doppler sy'stem where far field operation is

required at present (although a focusing action should be possible).

TIle holographic technique provides excellent results, probably the

best that have .been produced by any method so far, but the apertures used

are relatively large and these are essentially part of the image conversion

section of the system, rather than part of the processing technique. If

the same apertures were used with the other processing methods, then they

would also be expected to produce comparable results. The imaging system

is more flexible than that provided by a lens, since the field of view

may be varied at will. This advantage is heavily outweighed by the long

processing time required for the scanning and photographic copying.

The most inherently flexible technique is that provided by computer

processing, and if a real time on-line system is considered, all imaging

r.equirements could be provided for. There are problems, however, one of

computation time increases approximately as the square of the aperture

dimensions. Another problem, which is shared with all systems where the

aperture is scanned before processing takes place, is that operation with

short pulses in conjunction with large apertures necessarily leads to a

long data collection time. A lens, on the other hand, does not have this

problem because processing of all the image points is effectively done in

parallel, i.e. at the same time.

The scanned array Doppler system has the same correlation between

resolution and aperture. Although an inherently very simple processing

method, in its present form it suffers from lack of sensitivity since only

one element of the array is acting as a receiver at anyone time. It

should be possible to overcome this, but at the expense of increased

complexity. Much further work is required to be done investigating this

technique before a proper asse.ssment on its qualities, or otherwise, can

be made.'

To sunnnarise, it wou'ld seem that lenses and mirrors offer the simplest

and most effective method of image construction for applications where

immersion is possible and where rapid changes of the field of view are not

required. Where flexibility in conjunction with reasonable speed is

required, computer processing offers distinct advanta8es, and where

4. The Solid-State Converter

4.1 Converter DevelopmeE!

By 1968 the highly reliable and relatively high frequency silicon

planar transistor was fast becoming a readily availabi~ and inexpensive

component. This fact and the expectation of a rapid advance in

integra-ted circuit technology, particularly in fabrication techniques, decided

consideration should be given to the feasibility of us:i.ngtransistor

ana-logue gates as a means of sampling the electrical signals developed on 5 the back face of a piezoelectric transducer. It had been shown by Smyth

and others that electron beam scanning of a piezoelectric plate transducer

was currently the most sensitive method of conversion, so that the

sensi-tivity of the proposed method was confidently expected to be at least equal

to and probably better than this. A ten element bipolar transistor (2N706)

scanned transducer array was constructed and results from underwater tests

with this were very encouraging. As a consequence of this initial success

a complete ten by ten transducer matrix, scanned by bipolar transistors

with shift register control was built. Full details, including results, of

this converter are given ill Appendix A. This simple system proved to be

highly reliable and within the limits of the inherent picture "grain" size

(0')' inch square on transducer plate), produced easily recognisable

pic-tures of simple objects. However, it was clear that serious image

de,gracl.a-don was being caused by the capacitive coupling of signals across cut+off

transistors. This cross-talk problem has been virtually overcome :tn the

present converter by the use of field effect transistors which possess a

very low OFF capacitance (typically around 0'1 pf) between their input and

output terminals. The following sections provide details of the design,

11111111111111111111

11111111111111111111

111111111111111111.

11111111111111111111

111111111111111111111

••••••••••

_{11111111111111111111}

11111111111111111111

••••••

11•••

111111111111111111.

~t-v ':)

o

0. ..:J Z V -Cl) lIJ

r

l"

r.

-

.

.

(1 I.lI

i-ll)

.

a

V) ~ lL.I ~

z

er:

uJ

t- _{Cl 0} ~

I..L.. tU

-

W 0 _j

::t: _::>

...

_1Ll

tI)

\!1

_cr

..J

9

<t

~ "'Z _ul

<C

_:i

U

z

:z

o, :>

0 <t:

_'2.

0

!:!

_«

U')

ex:

l-:

tI)

-z.

0 _l1.I

«

:l: le. 0: ILl

I-~ ~ Cl)

---~t-

.-w

L

• S3J..\f-9 ~NI~3"d IJ...,olN iV?IJ..'tl"3A

..

j_

--{l

. ):f31.S'.g'3~

•

.L.::fIHS , \l-J'.L~" '3-9U'.1..S ,I

I

-v,

'--0

-s

uJ

\-(/) )0-ul tf)

='

~ _J 0 ~ _J

_V

et: 7:11) _0:'

CCl-r'

.l~

v

~\-

OJ

z

:>

_J

§O

l1.J

er

_.0:

0

:c

ut

l-a::

w

>

z, 0

uJ

UI ::>

e

c-il

0 -' t-<=, • Zo..

t

<\-z::>

UJ

o 0 er:

~ ';)

~

e

0

Lt:

Figure 4.3(a)

Figure 4.3(b)

Typical converter output-transmitter off.

. I I I I I

4.2 General Principles of the FET Scanned Converter

The transducer, a lead-zirconate-titanate disc, has its front face

(Le. that upon which the ultrasonic waves impinge) completely silvered

and its rear face partially silvered to form a ten by ten matrix of one

hundred equal area but electrically isolated squares (figure 4.1). The

electrical signals ~rom the one hundred effective transducer elements so

formed are each passed through an FET analogue gate, their outputs being

commoned in ·such a way as to form ten separate lines representing the ten

horizontal lines of the matrix (figure 4.2). A ten stage shift register

is connected to these gates so that a single "0" shifted through the stages

causes vertical columns of gates to be successively opened. Thus a

hori-zontal scan of the transducer matrix is performed, but the output appears

on ten lines and vertical mUltiplexing has yet to be carried out. These

ten lines are then vertically scanned by a further ten FET gates controlled

in the same manner by a second ten stage shift register so that after each

horizontal scan is completed, the vertical scan is shifted one line down.

The resulting output becomes a single line on which the complete trans ducer

matrix signals in a multiplexed serial form appear. Since the analogue

gates do not perform any demodulation process, this output signal is an

amplitude modulated carrier at the ultrasonic frequency, the peak amplitude

of which at any instant is proportional to the ultrasonic intensity on the

transducer element being sampled at that instant (see· figure 4.3).

4.3 Constructional Details

The FET converter matrix was constructed in two parts, the transducer

assembly and the FET scanning matrix, these being electrically connected

before final encapsulation. A third part, the control circuitry, is remote

from the converter matrix and twenty-three wires connect this to the first

ULTRASONIC CAM EllA 10 le 10 HT MATRIX

C')

0

0 0 0 0 0 0

• • • •

_--'01

• • •

0 G

or

0

•

•

•

•

•

•

•

•

• • 0 0 0 -0

• •

0 0 0

• 0

0 0

" • 0

•

0

·

o 0 0 0

•

-

_-0

• •

..

0

• •

w •

0 G •

•

•

• • • •

_{o to·}

0

"

•

•

0

• •

0

" 0

•

0

•

•

•

• •

• 0 .. 0 • 0 0 _-0

~ _• 0

•

0 o 0 • • G

•

• •

0 0 e

•

e 0 0 0

"

• • • • •

• Q 0

•

0

•

•

0 •

" • 0

..

0

• •

0

•

e

•

0 0 " 0

0 0 0 0 0 0

• •

0 0 ....l_ ₀

0

•

•

c

..

•

•

0 •

•

•

" 0

"

•

0 0 0

• •

u

-c

•

Cl 0 • 0 0 0 0

:1

•

•

• •

•

• •

•

0

• • • • • •

_"

•

e

0 ₀

---.-

-,

The transducer assembly consists of a PZT5 (manufactured by Vernitron

Ltd.) disc of 50 mm diameter, thickness resonant at 1MHz. Both faces

have a double layer of fired on silver, but, by request, the special

prim-ing layer between ceramic and silver had been omitted so that etching

cou.ld be carried out. Using a photo-resist technique a pattern of elect··

rodes (figure

4.1)

was etched with a ferric nitrate etchant. The pattern

includes a small circular electrode to be used as a transmitter for pulse-:

echo work. The main area consists of a square matrix of one hundred

ele-ments, each 80 mil by 80 mil with overall dimensions of one inch hy one

inch. Low temperature silver loaded solder and a temperature controlled

iron enabled one inch long unscreened 32 SWG wires to be attached to each

matrix electrode without fear of exceeding the Curie temperature. To

in-sure adequate mcchanic al strength after soldering a potting epoxy resi.n was

applied to a depth of approximately

!

inch around the wires.

The FET scanning matrix was constructed (figure 4.4) using a

multi-layer printed circuit (figure 4.5) to form the electrical int.erconnections

between the 110 FET's. The 100 FET source leads to be connected to the

transducer electrodes were arranged to project on the under surface so thnt

by positioning the scanning matrix over the transducer assemLly the

appro-priate \li.reson one could be soldered to those on the other.

Finally, when a supporting frame and all the necessary control and

signal wires had been attached, the combined assembly was encapsulated in

silicone rubber (Silastoner 70). The final unit measures approximately

90 x 115 x 50 mm (see figure 4.6).

4.4 The Transducer

if,4. 1 Choice of Transducer Material

whether the choice, if unrestricted, wou ld have re.mained "Tith-the

lead--zirconate-titanate ferro-electric ceramic PZT-SA used. Three properties

of the material are important in relation to its application to the image

converter. Namely:

(i) its Curie point temperature should preferably

allow soldering direct to the transducer metal

coating,

(ii) it should have a high receiving sensitivity coupled with a low output resistance in relation

to the electrical load,

(iii) the lateral spreading of waves wi thin the

trans-ducer should be as small as possib le.

Piezoelectric canst. e33

C m-2

,_J~5Xl~~L

Table 4.1

0'17

Quartz

x cut.

Curie point temp.

°c

573

Receiving const. g33

10-3 V m N-1 57'8

Mechanical Q

,---,---'7T-4 PZT-5A PZT-5H PZT-7A PZT··

--

f---

-,--,-328 365 193 350 3

26'1 24'8 19'7 39'9

24-15'1 15'8 23'3 9'5 13'

500 75 65 600 > 1

--

---'---1

I

8

I

i

._--.1

00

I

8 P,-,

8

03 .

_j

Consideration of the various constants given t.n table fr.l for quartz and the available PZT materials would immediately exclude PZT-5H owing to its

low Curie point temperature. With the except.ion of PZT-8 it can be seen,

have high receiving constants and are therefore more sensitive. In this

respect quartz is clearly the best, with PZT-7A next. However it has

40

been suggested that a greater lateral spreading of the waves occurs 1n

materials which possess high mechanical Q's, and on tI:isbasis PZT-5A

would be the choice, PZT-5H already having been excluded. Further

invE:s-tigation of the lateral spreading phenomenon is needed but it is

under-stood to be closely connected with mode conversion to shear and surface

waves within the transducer material (and possibly also within the backing

material, if solid). The effective electrical source resistance of the

transducer is also an important consideration, since some finite lORding

due to the amplifier input resistance must be present. The piezoelec.tric

constant e33 determines the mechanical to electrical transformation ratio and calculations for both quartz and PZT-5A are given in section 4.4.2

following. They show quartz as having a source resistance more than 4

orders larger than that for PZT-5A.

4.4.2 Source Resistance of Tuned Receiver Element

The back face loading of the transducer elements is virtually

incal-culable since the depth of solder, araldite and silicone rubber is not

accurately kaown, and in any case varies from element to element. It:will

therefore be assumed that the element is symmetricalLy loaded. The mechanical resistance R is given by,

m

R

=

peA

m

4.1

-3

where p is the density of the wat er in kg In ,C is the velocity of sound

-1

in water in ms, and A is the transducer element cross-·sectional area in ffi2. The transformaticn factor 0::1 converts the mechanical resistance into the effective electrical source resistance, R ohms, so that,

.